posted 28. January 2003 23:35                   

quote:

---

Evolution is an immensely powerful creative process. From the intricate biochemistry of individual cells to the elaborate structure of the human brain, it has produced wonders of unimaginable complexity. Evolution achieves these feats with a few simple processes--mutation, sexual recombination and natural selection--which it iterates for many generations. Now computer programmers are harnessing software versions of these same processes to achieve machine intelligence. Called genetic programming, this technique has designed computer programs and electronic circuits that perform specified functions.

---

posted 29. January 2003 01:41                   

quote:

---

The essential insight is that trial and error may only operate within a given hypervolume—but it may never jump to a new, higher-order hypervolume. The unbridgeable gaps between hypervolumes correspond to the technical contradictions in TRIZ theory.

---

quote:

---

The most obvious way to optimize variable dimensional problems with evolutionary
algorithms is to use variable length chromosomes for the EA, e.g. [3, 4]. This approach
requires the definition of new operators to increase or decrease the length of the internal
representation. Following the paradigm of natural evolution, these operators are
commonly modeled as gene deletion and gene duplication.

---

quote:

---

To achieve this, an evolutionary system which
identifies successful combinations of low-level, (basic) genes and combines them into
higher-level, (complex) genes is presented. Genes evolve in ever-increasing complexity, thus
encoding a higher number of the original basic genes. This results in a continuous
restructuring of the search space, allowing potential successful solutions to be found in
much shorter search time. This restructuring changes the landscape by changing the
probabilities of particular parts of the space being located.

---

quote:

---

One of the well-established notions related to creative designing processes is that an important means of characterizing them is to determine whether they have the capacity to expand the state space of possible designs - exploration (Gero, 1994).

---

posted 29. January 2003 20:50                   

quote:

---

The bottom line is that Darwinian processes cannot overcome technical contradictions and hence never generate inventive solutions to problems.

---

quote:

---

Trends Biochem Sci 2000 Jun;25(6):261-5

Evolution of a metabolic pathway for degradation of a toxic xenobiotic: the patchwork approach.

Copley SD.

The pathway for degradation of the xenobiotic pesticide pentachlorophenol in Sphingomonas chlorophenolica probably evolved in the past few decades by the recruitment of enzymes from two other catabolic pathways. The first and third enzymes in the pathway, pentachlorophenol hydroxylase and 2,6-dichlorohydroquinone dioxygenase, may have originated from enzymes in a pathway for degradation of a naturally occurring chlorinated phenol. The second enzyme, a reductive dehalogenase, may have evolved from a maleylacetoacetate isomerase normally involved in degradation of tyrosine. This apparently recently assembled pathway does not function very well: pentachlorophenol hydroxylase is quite slow, and tetrachlorohydroquinone dehalogenase is subject to severe substrate inhibition.

---

posted 30. January 2003 02:15                    

Yersinia,

(By the way, are you aware (I'm sure you are) that the word that usually comes right after your name is pestis (the organism responsible for Boubonic plague)? I know you don't really want to be known as a pest, but every time I read your name my mind fills in the next word and thats the unconscious association that comes to mind. sorry.)

On to your example. Both PCP degradation (which you described quite nicely) and other variants such as nylon-digesting bacteria are excellent examples of the power of Darwinian processes to search out the nooks and crannies of functionality of a given hyperspace of possibility. Both PCP degradation and nylon digestion are accessible by simple variation from a pre-existing starting point, namely the genes that were mutated to give rise to the PCP resistance or nylon digestion cabapilities respectively. In terms of sequence space both the "solutions" were quite close to the starting position (even though in terms of function they are drastically different). Mutations just had to "step over" to the solution by a simple point mutation (in the PCP example, there is cross-reactivity between the mutated pcpC enzyme and its original substrate, so likely the enzyme hasn't really changed that much). The only point of I have disagreement with your account is that you want to argue it's a "multi-part solution"--in reality, it's a single part that got added to a pre-existing pathway to impart novel functionality. And that single added part wasn't really changed that much (though once basic function was achieved, the whole pathway could be optimized by further "tweaking" to find the fitness peak). In all, it's a routine solution to a routine problem--there is nothing inventive about the solution (again, if you disagree, please demonstrate, using the TRIZ definition of inventiveness, to show how this was an inventive solution).

Inventive solutions always involve changing the very parameters of the evolving system. In other words, inventive solutions alter something that is fundamental to the evolutionary process itself (and hence is not accessible for evolution to alter). In the biomorphs example, the addition of a new "gene" with the novel function of coding for color was a prime example. This also explains where Rex makes a mistake--by assuming that gene duplication allows for inventive solutions to be found. What if Dawkins had just duplicated the gene for, say, branching depth? Or some other gene? He would just get a duplicate gene that would be activated at the same time as the "parent" gene and would do exactly the same thing as the "parent" gene (hopefully it wouldn't cause a syntax error or crash the program). There is a pre-defined set of "rules" that determine how the genes are interpreted by the program--and that system had to be altered by adding in the code that reads the "color" gene and interprets it properly as a color on the screen. Once the code is in place to read the "color" gene, that portion of code isn't "mutated" any more, but is used to interpret the parts that do evolve--the actual value of the color gene. Indeed, if the wrong portion of code is mutated (like the "rules" that outline how genes are interpreted), chaos would result as the program breaks down.

In my previous discussion of my paper, I said

quote:

---

...the process of embryonic development is a set of basic parameters which define the hypervolume in which the organism resides. Mutations to the genome serve to move about within that hypervolume (by their effect, "filtered" through development), but they do nothing to alter the fixed paramters of embryonic development in which the organism resides. (Notice I'm not saying that no mutations ever affect embryonic development; surely some mutations cause it to go badly awry or to terminate early, etc. But these do not alter the hypervolume that the organism exists in; they only mess up what is already established.)

Gene regulation is another factor in establishing a hypervolume of possibilities. Does a particular gene mutation kill the organism or create a favorable mutant? It depends on how it impacts embryonic development and also the expression of other genes. It is the logical interrelationship of genes that determines what changes are viable and hence selectable. Perhaps we can envision the process of embryological development as the interactions of genes through time: a program of certain interactions that give rise to a certain type of organism.

---

posted 30. January 2003 06:17                   

Who says I have to be Y. pestis (aka bubonic plague, which they apparently lost a bit of from a lab in Texas)? There are other kinds of Yersinia...



Before I attempt to answer your question, John, I would like to try and narrow down the exact, final, definitive definition of "inventive solution" in TRIZ-terms. The best I can seem to find are terms like "non-routine" and "resolving a technical contradiction", which help a bit but are difficult to apply.

Is there perhaps an authoritative quote you could post that would constitute the best-available definition?

Perhaps (especially since you focused on morphological evolution in your reply) answering the following would help:

Regarding morphology/development, would the following set of stages, if evolved through, constitute an "inventive" or "routine" solution:

1) Start with segmented metazoan.

2) Duplicate a segment (e.g. the corresponding Hox gene is serially duplicated). Critter now has an extra segment somewhere in the middle (say, 5 instead of 4) and corresponding pair of legs, etc., that go with the segment.

3) Repeat step 2 a number of times (e.g., selection for larger body size retains these duplications)

4) Once there are a fair number of segments, mutation and selection modify one or several of the more forward pairs, e.g. to improve prey capture or food-chewing or substrate/mate clasping (large number of possibilities here, lots of arthropods have these kinds of specializations).

Below is a slightly less abstract case (in arthropods, but dealing with the origin of a novel structure not from legs but from another structure). Would we have a novel or routine kind of solution here?

quote:

---

Proc Natl Acad Sci U S A 2002 Apr 16;99(8):5498-502

[which is free online I think]

Origin of a complex key innovation in an obligate insect-plant mutualism.

Pellmyr O, Krenn HW.

Evolutionary key innovations give organisms access to new ecological resources and cause rapid, sometimes spectacular adaptive radiation. The well known obligate pollination mutualism between yuccas and yucca moths is a major model system for studies of coevolution, and it relies on the key innovation in the moths of complex tentacles used for pollen collecting and active pollination. These structures lack apparent homology in other insects, making them a rare example of a novel limb. We performed anatomical and behavioral studies to determine their origin and found evidence of a remarkably simple mechanism. Morphological analyses of the tentacles and adjacent mouthparts in pollinators and closely related taxa showed that the tentacle appears abruptly in female pollinating yucca moths. Several morphological synapomorphies between the galeae, which constitute the characteristic lepidopteran proboscis, and the tentacle suggest that the tentacle evolved quickly through expression of the genetic template for the galea at an apical growth bud on the first segment of the maxillary palp. Behavioral data indicate that tentacle and proboscis movements are controlled by a shared hydraulic extension mechanism, thus no new mechanism was needed for tentacle function. Known developmental paths from other insects can explain the origin of this sex-specific key innovation in a few steps.

---

posted 31. January 2003 03:14                   

quote:

---

You can see how mutating certain key genes would simply destroy the developmental process...Since the developmental process consists of a complex system of interacting genes, small changes cannot re-wire the entire system or produce the system de-novo. This is why I argue that fundamentally, the developmental process is fixed

---

quote:

---

Furthermore, how would one make major alterations in such a gene network, once it is established?

---

posted 31. January 2003 05:14                   

quote:

---

So the origin of developmental programs should correspond to an inventive change in biology. It's where the "rules" came into existence (or were re-defined from what previously existed). Part of the problem for an evolutionary account is that there is no way to have a half-set of rules that govern something like embryonic development. The whole thing is deeply teleological; the entire process points toward an end goal: a functioning, viable adult organism.

---

posted 31. January 2003 10:40                

quote:

---

One of the points that came up on ARN was that evolutionary algorithms can be very good at squeezing every bit of functionality from a given hypervolume of possibilities (i.e, it's very good at sifting variants). We humans do the inventive work by setting up that hypervolume, but the program is good at working within that volume to pull out solutions that we may not be able to come up with directly (usually because they are intractable, see the SELEX example from my paper). So yes, in a sense these programs are "inventive" in an intuitive way. But TRIZ theory gives a definition of "inventive" which is more rigorous: an invention is made when an contradiction is overcome--or, to put the same thing another way, when the boundaries of a hypervolume are shifted such that new solutions are reachable. This means that even a lot of human inventions aren't strictly inventive, and that was one of Altschuller's points--that many patents are issued for trivial, routine solutions that don't qualify as inventions. None of these routine solutions overcome a contradiction (routine solutions are the ones found within a hypervolume, while inventive solutions re-engineer the hypervolume to allow new possibilities).

---

• Basics of logic elements (generic logic functions, storate elements of flip flops)
• Concept of arbitrary connections thereto
• Specified goal

posted 31. January 2003 21:57                   

quote:

---

Whiting says, however, that while wing re-evolution may seem unlikely to insect researchers, the basic idea of switching regulatory genes off and on is well accepted. Even a single gene can sometimes switch on the growth of a complex structure - studies indicate that a master gene called Pax-6, for example, might control the development of eyes in all creatures that have them.

http://www.newscientist.com/news/news.jsp?id=ns99993269

---